JP2011066592A - Coding mode selecting method, coding mode selecting device and coding mode selecting program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To adaptively select and switch frame coding or field coding in video coding in a small amount of operation processing. <P>SOLUTION: Each of a frame image with which fields are combined for an input image and a field image in which a field is divided by a field division section 10, is subjected to low-pass filtering in a vertical direction by a low-pass filter processing section 141, and a coding mode is selected from error information such as PSNR between video before filtering and video after filtering, without motion detection. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a method and apparatus for adaptively selecting frame coding and field coding in units of pictures or blocks in moving picture coding.

H. In video encoding standards such as H.264 / AVC, there are the following four encoding methods in order to realize highly efficient encoding.
(1) Frame encoding for encoding one picture as one frame,
(2) Field coding for coding one picture into a top field and a bottom field,
(3) Picture level field frame adaptive coding in which field / frame coding is switched for each picture to perform coding.
(4) Macroblock level field frame adaptive encoding in which field / frame encoding is switched in units of macroblock pairs.

  The frame coding method (1) is suitable for coding a moving image with little motion, and the field coding method (2) is suitable for coding a moving image with motion. , A scene with little motion and a scene with motion are mixed in one moving image sequence. For this reason, an encoding method for adaptively switching field / frame encoding at the picture level (3) is prepared.

  In addition, in a picture, a region with little motion and a region with motion are mixed, and in order to cope with this, field / frame coding is performed at a predetermined region level (macroblock level) in (4). There is an encoding method for adaptively switching between.

  In general, encoding can be improved by using the adaptive switching encoding method (3) or (4) and selecting a method that increases the encoding efficiency for each picture or macroblock.

  On the other hand, a field / frame coding switching method using motion information as described in Patent Document 1 and Patent Document 2 is known. However, there is a problem that derivation of motion information, that is, motion detection processing has a large calculation amount and is affected by motion detection error. The calculation amount is as light as possible, and the influence of errors included in information used in the switching processing is small. Processing is desirable.

  Also, which of (3) and (4) is highly efficient depends on the picture. For example, when coding is performed on a picture with motion as a whole, the picture field coding selected in (3) is the most efficient, but when (4) is used, the header information increases, so the efficiency is increased. Will decline. On the other hand, when encoding a locally moving picture, the frame or field encoding selected for each region in (4) is the most efficient, but if (3) is used, the prediction error is Since it becomes large, encoding efficiency falls. From the above viewpoint, it is desirable to combine (3) and (4) and adaptively switch according to the picture.

  FIG. 8 is a diagram illustrating an example of a moving picture coding apparatus using a conventional coding mode selection method. Hereinafter, an example of a method for adaptively selecting and switching the conventional coding mode will be described with reference to FIG.

  First, when original image data is input, the field dividing unit 50 divides the field image into a top field image and a bottom field image. The encoding mode determination criterion calculation unit 51 receives the frame image of the original image data and the two field images divided by the field dividing unit 50, and divides each image into coding unit blocks by the block dividing unit 511. To do. For each of the divided frame block image and field block image, the frame / field motion detection unit 512 calculates motion information and an evaluation function value based on a predetermined evaluation function such as mean square error (MSE). Further, a difference Dif between the evaluation function value of the frame image and the evaluation function value of the field image is calculated for each block. Motion information such as motion vectors is stored in the motion vector storage memory 56.

  The block number counting unit 52 counts the number of blocks with a difference Dif of the evaluation function value equal to or greater than a threshold value for one picture, and calculates a ratio R of the number of blocks to the total number of blocks in the frame. If the ratio R is within a certain range, the block unit field / frame encoding selection unit 54 is selected. Otherwise, the picture unit field / frame encoding selection unit 53 is selected.

  The picture unit field / frame coding selection unit 53 selects the coding mode of picture unit frame coding or picture unit field coding according to the value of the ratio R, switches the switch SW, and encodes the original image data. 60 pictures are transmitted to the frame unit 602 or the picture field encoding unit 603. The picture unit frame encoding unit 602 performs frame encoding on a picture basis using the motion information stored in the motion vector storage memory 56. Similarly, the picture unit field encoding unit 603 performs frame encoding in units of pictures using the motion information stored in the motion vector storage memory 56.

  The block unit field / frame coding selection unit 54 determines the value of the ratio R, and performs block-level adaptive coding when a large motion portion and a small motion portion are mixed in the picture. Therefore, depending on whether or not the difference Dif between the evaluation function value of the frame image and the evaluation function value of the field image is equal to or larger than a certain threshold value, field encoding or frame encoding is performed in units of blocks (macroblock pairs). The coding mode selection information is stored in the coding mode storage memory 55. At this time, the original image data is sent to the block unit encoding unit 601, and the block unit encoding unit 601 stores it in the motion vector storage memory 56 according to the encoding mode stored in the encoding mode storage memory 55. Using the existing motion information, encoding is performed on a block-by-block basis in the frame encoding mode or the field encoding mode.

  In the above prior art, the frame / field motion detection unit 512 must always execute motion detection processing for all frame images and field images before selecting an encoding mode. Is big.

JP-A-5-308627 Japanese Patent Laid-Open No. 5-95545

  The present invention has been proposed to solve the above problems, and its purpose is to adapt frame coding / field coding in moving picture coding with a small amount of processing computation based on the frequency characteristics of pictures. It is to make it possible to perform switching selection automatically, that is, to select an appropriate encoding mode in a short time.

  In order to solve the above-described problems, the present invention provides a frame image and a field dividing unit that combine fields for an input image instead of using an evaluation function value such as a mean square error (MSE) calculated at the time of motion detection for mode selection. For each of the top field image and the bottom field image obtained by dividing the field in Step 1, a low-pass filter is applied in the vertical direction, and the PSNR (Peak Signal to) between the image before filtering and the image after filtering is applied. The main feature is to select an encoding mode from error information such as noise ratio.

  That is, according to the present invention, in moving picture coding in which frame coding and field coding are adaptively selected and coded for interlaced video, the input picture is a field picture (both field picture and top field). The input image is subjected to filter processing in frame picture units and field picture units with a predetermined bandwidth. In addition, the input image is divided into blocks each composed of a plurality of pixels in a coding unit, and error information (for example, PSNR) between the filtered frame picture (also referred to as a frame image) and the input image is calculated in units of blocks. The error information between the filtered field picture and the field picture of the input image is calculated in units of blocks, and the difference between the block unit error information between the frame picture and the field picture is calculated as a coding mode determination criterion. Using the calculated coding mode criterion, one coding mode is selected from among coding modes including any of picture unit frame coding, picture unit field coding, block unit frame coding, or block unit field coding. Select a mode.

  In the above invention, in the process of selecting the coding mode, the number of blocks in which the difference in block unit error information between the frame picture and the field picture is equal to or greater than a predetermined threshold is counted, and the total number of blocks counted. A ratio with respect to the number is calculated, and a coding mode of picture unit frame encoding or picture unit field encoding is selected from the ratio.

  Further, in the above invention, in the process of selecting the coding mode, the block unit frame coding is performed by comparing the difference between the block unit error information between the frame picture and the field picture and the predetermined threshold value on a block basis. Alternatively, any coding mode of block unit field coding is selected.

  In the above invention, it is also preferable to increase or decrease the threshold used for selecting the encoding mode according to the quantization parameter, the target code amount, or the generated code amount.

  According to the present invention, frame coding / field coding in moving picture coding can be adaptively switched and selected based on the frequency characteristics of a picture and with a small amount of processing computation. Encoding can be performed.

It is a figure which shows the structural example of the moving image encoder which uses the encoding mode selection method which concerns on the 1st and 2nd example of this invention. It is a figure which shows the structural example of an encoding mode determination reference | standard calculation part. It is a figure which shows the structural example of a block unit encoding part. It is a figure which shows the structural example of the frame unit for a picture unit. It is a process flowchart of the 1st example. It is a processing flowchart of the 2nd example. It is a figure which shows the structural example of the moving image encoding apparatus using the encoding mode selection method which concerns on a 3rd example. It is a figure which shows an example of the moving image encoder which uses the conventional encoding mode selection method.

  Embodiments of the present invention will be described below with reference to the drawings.

<Configuration example of moving image encoding device>
FIG. 1 is a block diagram showing a configuration example of a moving picture coding apparatus using a coding mode selection method for adaptively selecting frame coding and field coding in a first example and a second example described later. 1 includes a field dividing unit 10, an encoding mode determination criterion calculating unit 14, a block number counting unit 15, a picture unit field / frame encoding selecting unit 16, and a block unit. A field / frame coding selection unit 17, a coding mode storage memory 18, and a coding unit 20 are provided.

  The encoding mode determination criterion calculation unit 14 includes a frame image 11 obtained by combining the fields of the original image, a top field image (denoted as 1st field image in the drawing) 12 and a bottom field image (in the drawing, indicated in the drawing). , 2nd field image) 13, a low-pass filter processing unit 141 that performs low-pass filter processing on each of the 13, a block dividing unit 142 that divides these images into blocks composed of a plurality of pixels of coding units, and error information A difference information calculation unit 143 is provided. The error information / difference information calculation unit 143 performs processing for calculating error information (for example, PSNR) between the filtered frame image and the input image in units of blocks, and the error between the filtered field image and the input image field image. A process of calculating information in units of blocks and a process of calculating a difference in block unit error information between the frame image and the field image as an encoding mode determination criterion are performed.

  The encoding unit 20 includes a block unit encoding unit 200 that performs encoding by switching between field encoding and frame encoding in block units during block level adaptive encoding, and a picture unit in picture unit frame encoding mode. A picture unit frame encoding unit 300 that performs frame encoding in units and a picture unit field encoding unit 400 that performs field encoding in units of pictures in the picture unit field encoding mode.

  The apparatus shown in FIG. 1 differs from the prior art mainly in the encoding mode determination criterion calculation unit 14 and the encoding unit 20, and the encoding mode determination criterion calculation unit 14 performs motion detection with a large calculation load. The encoding mode determination criterion is calculated without performing this, and the encoding unit 20 is different from the prior art in that only encoding is performed by performing motion detection on a field image or a frame image necessary for encoding.

  FIG. 2 shows a configuration example of the encoding mode determination criterion calculation unit 14 shown in FIG. A low-pass filter processing unit 141 shown in FIG. 1 applies a low-pass filter with a predetermined bandwidth r to the frame image 11, and a low-pass filter with a predetermined bandwidth r for the 1st field image 12. A low-pass filter processing unit 1412 that applies a low-pass filter and a low-pass filter processing unit 1413 that applies a low-pass filter having a predetermined bandwidth r to the 2nd field image 13 are configured.

  The block dividing unit 142 shown in FIG. 1 includes block dividing units 1421, 1422, and 1423 that divide the filtered frame filter image, 1st field filter image, and 2nd field filter image into blocks, respectively.

  The error information / difference information calculation unit 143 illustrated in FIG. 1 includes an SNR calculation unit 1431 that calculates a PSNR between the frame image and the original image in block units as error information, and a PSNR between the 1st field image and the original image. SNR calculation unit 1432 for calculating SNR, SNR calculation unit 1433 for calculating PSNR between the 2nd field image and the original image, and an average for calculating the average value of the output of SNR calculation unit 1432 and the output of SNR calculation unit 1433 The value calculation unit 1434 includes a subtracter 1435 that calculates a difference between the output of the SNR calculation unit 1431 and the output of the average value calculation unit 1434 and outputs error difference information as a coding mode determination criterion.

  FIG. 3 shows a configuration example of the block unit encoding unit 200 shown in FIG. When the original image data is input, the block unit encoding unit 200 first divides the block in the frame mode from the encoding mode stored in the encoding mode storage memory 18 in the block dividing unit 201 or the block in the field mode. Decide whether to divide, and divide the original image data into blocks. Subsequent processing is the same as encoding processing using normal motion compensation. First, a motion vector is detected by the motion detection unit 202, and a prediction signal is generated by the prediction unit 203 from the motion vector and the local decoded image 208. To do.

  The subtractor 220 calculates a prediction residual signal that is the difference between the block-divided original image signal and the prediction signal, and the DCT transform unit 204 performs orthogonal transformation of discrete cosine transform on the prediction residual signal. . The quantization unit 205 quantizes the orthogonal transform coefficient and outputs it to the variable length coding unit 209 and the inverse quantization unit 206. The variable length coding unit 209 performs entropy coding on the quantization result. The multiplexing unit 210 multiplexes the encoded data of the encoding mode in block units, the encoded data obtained by quantizing the orthogonal transform coefficient, and the encoded data of the motion vector, and stores them in the buffer unit 211. The encoded data stored in the buffer unit 211 is output as an encoded stream. The control unit 212 controls the generated code amount in the quantization unit 205 from the generated code amount of the encoded data stored in the buffer unit 211 and the target code amount.

  The inverse quantization unit 206 inversely quantizes the output of the quantization unit 205, and the inverse orthogonal transform unit 207 performs inverse orthogonal transform on the orthogonal transform coefficient obtained by the inverse quantization to calculate a prediction residual signal. The result and the prediction signal are added by the adder 221 to generate a local decoded image 208, which is stored in a memory for use as a reference image in later prediction encoding.

  FIG. 4 shows a configuration example of the picture unit frame encoding unit 300 shown in FIG. The block unit encoding unit 200 shown in FIG. 3 performs block-level adaptive encoding that switches between frame encoding and field encoding on a block basis. However, the picture unit frame encoding unit 300 uses the entire current picture. In other words, frame coding is performed on all blocks in the picture. Therefore, unlike the multiplexing unit 210 in FIG. 3, the multiplexing unit 310 does not multiplex the encoded data of the encoding mode in units of blocks. The basic functions and operations of the other parts are the same as those of the block unit encoding unit 200 shown in FIG.

  The picture unit field encoding unit 400 shown in FIG. 1 performs the picture unit frame code shown in FIG. 4 except that the input field of the original image data is subjected to field encoding of the top field and the bottom field in units of pictures. Since the configuration is the same as that of the conversion unit 300, the description using the drawings is omitted.

<First example>
FIG. 5 is a processing flowchart of the first example of the moving picture encoding apparatus shown in FIG. In the first example, the block unit encoding unit 200 and the picture unit field encoding unit 400 are used as the encoding unit 20, and the picture unit frame encoding unit 300 is not used. That is, the first example is an example in which the picture unit frame coding mode is not selected.

  First, in step S100, original image data B_o is input, and in step S101, the field dividing unit 10 divides the original image data B_o into top field image data B_tfl and bottom field image data B_bfl.

  In step S102, the divided field images are input to the low-pass filter processing unit 141, and are subjected to filter processing with a predetermined bandwidth r to generate field filter image data B_tfl_fil and B_bfl_fil. On the other hand, the frame image data B_fr that is not divided is the original image data B_o itself, which is directly input to the low-pass filter processing unit 141, and subjected to filter processing with the same r as the bandwidth, so that the frame filter image data B_fr_fil Is generated.

  In step S103, the obtained filter image data and original image data are each divided into blocks by the block dividing unit 142. Subsequently, in step S104, error information (for example, PSNR) of the divided blocks is calculated by the error information / difference information calculation unit 143. Here, N is a block number given to the divided block, and block unit error information of top field filter image data, bottom field filter image data, and frame filter image data is P_tfl [N], P_bfl [N], and P_fr [ N]. Also, P_fl [N] = (P_tfl [N] + P_bfl [N]) / 2. A difference Db = P_fr [N] −P_fl [N] of these error information is calculated. That is, the error information difference Db is a value obtained by subtracting the PSNR of the field image (the average value of the PSNR of the top field image and the PSNR of the bottom field image) from the PSNR of the frame image in units of blocks. The fact that the difference Db is large means that the degree of image quality deterioration of the field image due to the filter processing is considerably larger than the degree of image quality deterioration of the frame image.

  In step S105, the number of blocks satisfying the following conditions is set to M in the block number counting unit 15.

Db ≧ Th1 (a)
Here, Th1 is a constant given as a threshold and varies depending on the input image size, the characteristics of the filter used, and the like. From the count result of M, R which is a ratio (ratio) of the block M satisfying the condition (a) with respect to the total number of blocks included in the picture is calculated.

R = M / total number of blocks In step S106, it is determined whether or not R is greater than a certain threshold value Tr. If R is smaller than the threshold value Tr, the picture unit field / frame coding selection unit 16 selects picture unit field coding. In step S107, the picture unit field encoding unit 400 performs picture unit encoding. The fact that R is smaller than the threshold value Tr is considered that degradation in image quality is considered to be small even when field coding is performed on all blocks of the picture to be coded. As will be described later, instead of PSNR as error information, for example, a mean square error (MSE) between the original image and the filtered image can be used. In this case, the magnitude relationship with the threshold value uses PSNR. The opposite is true. The same applies to other examples described below.

  If it is determined in step S106 that R is greater than or equal to the threshold value Tr, in the first example, the process proceeds to step S108, and block level adaptive encoding is selected. In the case of block level adaptive coding, the block level unit field / frame coding selection unit 17 operates, and the block unit error information and the frame filter image of the field filter image calculated by the error information / difference information calculation unit 143 in step S109. The block unit error information is referred to to determine whether or not the following condition is satisfied, and either block unit frame encoding or block unit field encoding is selected.

Db ≧ Th2 (b)
Here, Th2 is a constant given as a threshold value, and is different depending on the input image size, the characteristics of the filter to be used, and the like, similarly to Th1 in the condition (a). When this condition (b) is satisfied, block unit frame coding is selected, and information on the coding mode is stored in the coding mode storage memory 18. If the condition (b) is not satisfied, block-unit field encoding is selected, and information on the encoding mode is stored in the encoding mode storage memory 18.

  The block unit encoding unit 200 performs frame encoding on the target block in step S110 and performs block unit field encoding when a block unit frame encoding is performed on the encoding target picture using the encoding mode for each block. Sometimes, field encoding is performed on the target block in step S111.

  Hereinafter, the technique of the first example will be described using a specific example. The image size of the original image is 1920 × 1080, the filter bandwidth is r = 0.5 (a low-pass filter that cuts frequency components in half in the horizontal and vertical directions), the constants Th1 = 1, Th2 = 2, and Tr = 0. .8.

  First, the original image is divided into a top (1st) field and a bottom (2nd) field by the field dividing unit 10 to obtain a field image (1920 × 540). These are input to the low-pass filter processing unit 141, and filter processing is performed with a bandwidth of 0.5 to obtain field filter images B_tfl_fil and B_bfl_fil. On the other hand, when the original image is a frame image, it is directly input to the low-pass filter processing unit 141 and subjected to filter processing with a bandwidth of 0.5 to obtain a frame filter image B_fr_fil.

  Next, in the block dividing unit 142, each field filter image and frame filter image are divided into 16 × 16 and 16 × 32, respectively. Measure the PSNR for. The results are P_tfl [1] = 26 (dB), P_tfl [2] = 25 (dB),..., P_bfl [1] = 24 (dB), P_bfl [2] = 23 (dB),. Suppose that P_fr [1] = 27 (dB), P_fr [2] = 25 (dB),. Note that the number of divisions in the above case is 120 × 34 = 4080.

  Next, the block number counting unit 15 applies these to the condition (a). Assume that the number M of blocks that satisfy the condition (a) is 3000. In this case, since R (= 3500/4080) is less than Tr (= 0.8), the picture unit field / frame coding selection unit 16 selects picture unit field coding, and the coding unit 20 Encoding is performed and the process is terminated.

  On the other hand, when the number of blocks M is 3500, since R is greater than or equal to Tr, the following processing proceeds. The error information difference Db previously calculated by the error information / difference information calculation unit 143 is referred to, and the block unit field / frame encoding selection unit 17 applies the condition (b) for each block. Then, since the condition is satisfied in the first block, the block unit frame encoding is selected, and in the second block, the condition is not satisfied, so the block unit field encoding is selected. The encoding unit 200 performs block-level adaptive encoding and ends the process.

<Second example>
FIG. 6 is a processing flowchart of the second example of the moving picture encoding apparatus shown in FIG. In the second example, the coding unit 20 is used for the picture unit frame coding unit 300 in addition to the block unit coding unit 200 and the picture unit field coding unit 400. That is, the second example is an example in which a picture unit frame coding mode is also a mode selection target.

  The processing from step S201 to S207 in FIG. 6 is the same as the processing from step S101 to S107 in the first example shown in FIG. The second example is different from the first example in that two threshold values, Tr1 and Tr2, are prepared for comparison with the ratio R of the number M of blocks that satisfy the condition (a) with respect to the total number of blocks in the frame. In the example of 2, the comparison is made between R and Tr2 (step S208), and when R ≦ Tr2, the picture unit frame coding is performed on the target picture (step S209). Subsequent block level adaptive encoding processing is the same as in the first example described above.

  The threshold value Tr1 is the same as the threshold value Tr in the first example, and Tr2 is a threshold value that is larger than Tr1 (Tr2> Tr1).

That is, the encoding mode is selected as follows.
・ When R <Tr1:
The picture unit field / frame encoding selection unit 16 selects picture unit field encoding, and the picture unit field encoding unit 400 performs picture unit field encoding including field motion detection processing on the target picture.
・ When R> Tr2:
The picture unit field / frame coding selection unit 16 selects picture unit frame coding, and the picture unit frame coding unit 300 performs picture unit frame coding including frame motion detection processing on the target picture.
・ When Tr1 ≦ R ≦ Tr2:
Block level adaptive encoding is selected, and block unit field encoding / block encoding selection unit 17 selects block unit frame encoding and block unit field encoding by comparing error information difference Db and threshold value Th2 in units of blocks. I do. That is, if Db ≧ Th2, block unit frame encoding is selected for the block in the target picture, and if Db <Th2, block unit field encoding is selected for the block in the target picture. The block unit encoding unit 200 performs encoding.

<Third example>
FIG. 7 is a block diagram showing an example of a moving picture coding apparatus that adaptively selects frame coding and field coding in the third example. A quantization parameter adaptation unit 19 is further provided for the moving picture coding apparatus of FIG. The other parts are the same as those of the moving picture encoding apparatus described with reference to FIG.

  In the third example, by providing the quantization parameter adaptation unit 19, the thresholds Th1, Th2 and Tr (or Tr1, Tr2) used for selecting each coding mode are made variable according to the quantization parameter. . In the third example, that is, the threshold value used in the first example and the second example is changed in accordance with the quantization parameter, so that higher accuracy can be achieved.

  The quantization parameter adaptation unit 19 may adaptively change the threshold immediately before each mode selection, or may determine the thresholds collectively in the first part of the process.

  When the quantization parameter is small, the image is finely quantized. By finely quantizing, the image quality is improved accordingly. That is, if the quantization parameter is small, the generated code amount becomes large. Under this assumption, for example, the threshold is determined as follows.

Th1 and Th2 Since the field image is skipped by one line in the vertical direction, the correlation is lower than that of the frame image, and when a filter having the same bandwidth is applied in the vertical direction, the frame image is generally used. The SN ratio becomes worse than. Therefore, Th1 and Th2 are determined to appropriate values in consideration of the encoding efficiency. This value does not need to be changed depending on the quantization parameter because the influence is small even if the quantization parameter changes.

Tr (Tr1, Tr2) When frame / field encoding is switched in units of blocks, the encoding efficiency is higher than that in units of frames. However, it is necessary to add information for indicating which encoding is to be performed in units of blocks, and increasing this additional information may increase the overall amount of information and reduce the encoding efficiency. There is. Therefore, it is more efficient to set Tr larger (for example, Tr = 0.8). However, when the quantization parameter is small (the image quality is good), the amount of code increases, so the ratio of additional information to the total information amount decreases. Therefore, it is better to set Tr smaller when the quantization parameter is small.

<About other examples>
In the above example, the left side of condition (a) is the average error information value for the top field and bottom field, but since the correlation between the fields is high, it can be calculated using only one of the error information values. The same effect is produced. Further, although PSNR is used as error information, the same effect can be obtained when a mean square error or the like is used. Note that, in the comparison between the error difference information and the like in the mode selection and the threshold value, the magnitude relationship changes depending on what is used as the error information.

  Further, the bandwidth r used in the filter processing can take a value of 0 <r <1, and the filter characteristics differ depending on the value, and the filter constants differ accordingly, but the same effect is obtained. Filtering is performed with the same bandwidth in the vertical and horizontal directions, but the same effect can be obtained when using separate bandwidths in the vertical and horizontal directions, or when filtering is performed only in the vertical direction. .

  Further, although R is a block ratio, the same effect can be obtained when a table showing the relationship between the image size (resolution) and the number of blocks is used. In addition, whether or not to perform field-by-picture field coding is selected based on the result of passing through the block division unit 142, the error information / difference information calculation unit 143, and the block count unit 15, but the field filter image and the frame filter image The same effect can be obtained when the picture unit field coding is selected only from the error information (average value of the block unit error information). In addition, the quantization parameter adaptation unit 19 in the third example has the same effect when a target code amount, a generated code amount, or the like is used instead of the quantization parameter.

  Further, the filtering process in the low-pass filter processing unit 141 is a prefiltering process for the purpose of noise removal and improvement of coding efficiency, so that the prefiltering process and the field / frame selection process including the prefiltering process are performed. Can be performed in a batch, so it is highly consistent and can be highly functional without increasing the amount of computation, which is beneficial.

  The above encoding mode selection processing can be realized by a computer and a software program, and the program can be recorded on a computer-readable recording medium or provided through a network.

DESCRIPTION OF SYMBOLS 10 Field division | segmentation part 11 Frame image 12 1st field image 13 2nd field image 14 Encoding mode determination reference | standard calculation part 15 Block number count part 16 Picture unit field / frame encoding selection part 17 Block unit field / frame encoding selection part 18 Code | symbol Storage mode 19 Quantization parameter adaptation unit 20 Encoding unit 141 Low-pass filter processing unit 142 Block division unit 143 Error information / difference information calculation unit 200 Block unit encoding unit 300 Picture unit frame encoding unit 400 Picture unit field code Chemical department

Claims (9)

In a coding mode selection method in moving picture coding for adaptively selecting and coding frame coding and field coding for interlaced video,
Dividing the input image into field pictures of top field and bottom field;
A process of performing a filtering process on an input image in units of frame pictures with a predetermined bandwidth;
A process of filtering an input image in a field picture unit with a predetermined bandwidth;
A process of dividing an input image into blocks composed of a plurality of pixels in a coding unit;
A process of calculating error information between the filtered frame picture and the input image in units of blocks;
Calculating error information between the filtered field picture and the field picture of the input image in units of blocks;
Calculating a difference in block unit error information between the frame picture and the field picture as a coding mode criterion,
Using the coding mode criterion, a coding mode including any one of block unit frame coding or block unit field coding in picture unit frame coding, picture unit field coding, or block level adaptive coding And a method of selecting one encoding mode from among them. In the encoding mode selection method according to claim 1,
In the process of selecting the coding mode, the number of blocks in which the difference in block unit error information between the frame picture and the field picture is equal to or greater than a predetermined threshold is counted, and the ratio of the counted number of blocks to the total number of blocks And a coding mode selection method, wherein either one of picture unit frame coding or picture unit field coding is selected from the ratio. In the encoding mode selection method according to claim 1 or 2,
In the process of selecting the coding mode, the difference between block unit error information between the frame picture and the field picture is compared with the predetermined threshold value in block units, thereby comparing the block unit in block level adaptive coding. A coding mode selection method comprising: selecting a coding mode of either frame coding or block unit field coding. In the encoding mode selection method according to claim 2 or 3,
A coding mode selection method comprising a step of controlling a threshold used for selecting the coding mode according to a quantization parameter, a target code amount, or a generated code amount. In a coding mode selection apparatus in moving picture coding for adaptively selecting and coding frame coding and field coding for interlaced video,
Means for dividing the input image into field pictures of top and bottom fields;
Means for performing a filtering process on a frame picture basis with a predetermined bandwidth for an input image;
Means for performing filtering on the input image in units of field pictures with a predetermined bandwidth;
Means for dividing an input image into blocks composed of a plurality of pixels in a coding unit;
Means for calculating error information between the filtered frame picture and the input image in units of blocks;
Means for calculating error information between the filtered field picture and the field picture of the input image in units of blocks;
Means for calculating a difference in block unit error information between the frame picture and the field picture as a coding mode criterion;
Using the coding mode criterion, a coding mode including any one of block unit frame coding or block unit field coding in picture unit frame coding, picture unit field coding, or block level adaptive coding An encoding mode selection device comprising: means for selecting one encoding mode from among them. In the encoding mode selection device according to claim 5,
The means for selecting the encoding mode counts the number of blocks in which the difference in block unit error information between the frame picture and the field picture is equal to or greater than a predetermined threshold, and the ratio of the counted number of blocks to the total number of blocks And a coding mode selection device characterized by selecting a coding mode of picture unit frame coding or picture unit field coding from the ratio. In the encoding mode selection device according to claim 5 or 6,
The means for selecting the coding mode is configured to compare a difference between block unit error information between the frame picture and the field picture and a predetermined threshold value in block units, thereby comparing each block unit in block level adaptive coding. An encoding mode selection device that selects either one of frame encoding or block unit field encoding. In the encoding mode selection device according to claim 6 or 7,
A coding mode selection apparatus comprising: means for controlling a threshold used for selecting the coding mode according to a quantization parameter, a target code amount, or a generated code amount.   An encoding mode selection program for causing a computer to execute the encoding mode selection method according to any one of claims 1 to 4.

Images